Object recognition apparatus, object recognition method, and object recognition program

Abstract

An object recognition apparatus includes an image acquisition unit that acquires a captured image of a photographic subject, and a recognition processing unit that recognizes the photographic subject in the acquired image using a recognition dictionary. The recognition processing unit detects a target in the acquired image using a target recognition dictionary, detects a wheel at a lower part of the detected target using a wheel recognition dictionary, and reflects a result of the detection of the wheel in a result of the detection of the target. Thus, the object recognition apparatus can accurately detect a target such as a person and another vehicle.

What is claimed is: 
     
       1. An object recognition apparatus comprising:
 an image acquisition unit configured to acquire a captured image of a photographic subject; and 
 a recognition processing unit configured to recognize the photographic subject in the acquired image using a recognition dictionary, wherein the recognition processing unit
 detects a target in the acquired image using a target recognition dictionary, 
 detects an orientation of the target using the acquired image, wherein the orientation of the target is detected from an estimated direction in which the target travels, 
 selects a wheel recognition dictionary corresponding to the detected orientation, 
 detects a wheel at a lower part of the detected target using the selected wheel recognition dictionary, and 
 reflects a result of the detection of the wheel in a result of the detection of the target. 
 
 
     
     
       2. The object recognition apparatus according to  claim 1 , wherein when the recognition processing unit detects the orientation of the target using the acquired image, the recognition processing unit
 tracks the target using a plurality of the acquired images and estimates the direction in which the target travels. 
 
     
     
       3. The object recognition apparatus according to  claim 1 , wherein the recognition processing unit
 detects a person using a person recognition dictionary, which is the target recognition dictionary, 
 detects the wheel using the selected wheel recognition dictionary at a lower part of the detected person, and 
 decides, when the recognition processing unit detects the wheel, that the detected person is an occupant of the vehicle, and decides, when the recognition processing unit does not detect the wheel, that the detected person is a pedestrian. 
 
     
     
       4. The object recognition apparatus according to  claim 3 , wherein the recognition processing unit calculates a distance between a lower end of the detected wheel or a lower end of the detected person decided to be the pedestrian and a capturing unit that captures the image. 
     
     
       5. The object recognition apparatus according to  claim 1 , wherein
 the image acquisition unit acquires an infrared light image and a visible light image as the image, and 
 the recognition processing unit
 detects the person in the infrared light image using the person recognition dictionary, which is the target recognition dictionary, and 
 detects the wheel in the visible light image at the lower part of the detected person using the selected wheel recognition dictionary. 
 
 
     
     
       6. The object recognition apparatus according to  claim 1 , wherein the recognition processing unit
 detects a vehicle using a vehicle recognition dictionary, which is the target recognition dictionary, 
 detects a wheel at a lower part of the detected vehicle using the selected wheel recognition dictionary, and 
 calculates a distance to the detected vehicle based on a lower end position of the detected wheel. 
 
     
     
       7. The object recognition apparatus according to  claim 6 , further comprising a display control unit configured to control a display unit to display the acquired image, wherein the display control unit controls the display unit to display a detection rectangle surrounding the detected vehicle in such a way that the detection rectangle includes the detected wheel. 
     
     
       8. The object recognition apparatus according to  claim 7 , wherein when the display control unit controls the display unit to display the detection rectangle,
 when the recognition processing unit detects the vehicle, the display control unit controls the display unit to display the detection rectangle surrounding the vehicle, and 
 when the recognition processing unit detects the wheel, the display control unit corrects the detection rectangle so that it includes the wheel and then display the detection rectangle. 
 
     
     
       9. The object recognition apparatus according to  claim 7 , wherein the display control unit controls the display unit to display the calculated distance. 
     
     
       10. An object recognition method comprising:
 an image acquisition step of acquiring a captured image of a photographic subject; and 
 a recognition processing step of recognizing the photographic subject in the acquired image using a recognition dictionary, wherein the recognition processing step comprises:
 detecting a target in the acquired image using a target recognition dictionary; 
 detecting an orientation of the target using the acquired image, wherein the orientation of the target is detected from an estimated direction in which the target travels; 
 selecting a wheel recognition dictionary corresponding to the detected orientation; 
 detecting a wheel at a lower part of the detected target using the selected wheel recognition dictionary; and 
 reflecting a result of detecting the wheel in a result of detecting the target. 
 
 
     
     
       11. A non-transitory computer readable medium storing an objection recognition program for causing a computer to execute:
 an image acquisition procedure for acquiring a captured image of a photographic subject; and 
 a recognition processing procedure for recognizing the photographic subject in the acquired image using a recognition dictionary, wherein the recognition processing procedure comprises:
 detecting a target in the acquired image using a target recognition dictionary; 
 detecting an orientation of the target using the acquired image, wherein the orientation of the target is detected from an estimated direction in which the target travels; 
 selecting a wheel recognition dictionary corresponding to the detected orientation; 
 detecting a wheel at a lower part of the detected target using the selected wheel recognition dictionary; and 
 reflecting a result of detecting the wheel in a result of detecting the target.